1. Field of the Invention
A service model, integrated system, and method enabling a service provider to deliver an integrated web hosting and content distribution service offering, which provides assured service levels, regardless of whether the customer's web site is served by the service provider's hosting center, the service provider's content distribution network (CDN), or by a third party CDN.
2. Description of Background Art
Traditionally, content distribution network operators serve all-or-nothing for an area of a customer enterprise's web site; i.e., the enterprise operates its own web site for serving the content, or the enterprise outsources a name space (hostname) to a CDN operator to manage. There are currently some hybrid approaches offered by service providers. These include offering an enterprise needing web hosting a special price to include content distribution automatically as a part of a package, or offering the enterprise the ability to turn on a CDN as an “insurance policy” to back up an overloaded server. For an enterprise or organization with unpredictable loads, this can lead to unpredictability in cost and performance.
In current common business models, web hosting and content distribution networks (CDNs) are distinct. A web hosting service provider runs an “origin server” that is the ultimate site where requests for web content from client machines are handled. As shown in FIG. 1, the web hosting service provider may itself have a cluster of servers front-ended by a content-aware switch Sw, and serves requests from client machines, without any inherent content distribution network.
FIG. 2 shows a DNS-based CDN, and illustrates client machines C, local domain name server DNS, caches 2, as part of the IP network. The CDN can cache web content closer to the client machines C of the end users of its customers, thus improving performance and scaling to a larger number of customers. CDNs usually charge a premium for this service, partly due to the added cost of the infrastructure (additional caches 2), and primarily because enterprises are willing to pay a premium for more predictable operational performance.
There are two common models for using a CDN. Some content distribution providers require customers to add special coding to their HTML pages, to rewrite embedded links to retrieve objects exclusively from caches under its own operational control and ownership. On the other hand, other CDNs, including AT&T's CDN, for example, outsource a domain (for example images.company.com), to the CDN, making the CDN “authoritative” for where that content is served. The CDN resolves that hostname to an IP address for a cache (or cluster of caches) it serves.
An extension of the CDN model is something called “Content Internetworking”, a.k.a. “Content Distribution Internetworking” (CDI), as in FIG. 3. With CDI, a CDN can have another entity, such as a second CDN B, or third CDN C, resolve the name to an IP address in its own service. In this example, (1) a user using a client machine C goes to a customer local DNS server DNS1, which (2) sends a message to a CDI DNS server. That CDI DNS server returns a response (3) to the customer's local DNS server DNS1, which may cause it to make a subsequent DNS request (3′) of a DNS server at a second CDN (B in this case) getting a response (3″). The customer's local DNS1 (4) returns an IP address to the client machine C, which retrieves the data from either CDN B (5, 6) or CDN A (5′, 6′).